I currently serve as an Applied Scientist at Amazon AGI (Artificial General Intelligence), specializing in the development of audio foundation models within a multi-modal framework. Prior to my tenure at Amazon, I dedicated five years as a Staff Machiene learning Researcher at Qualcomm AI Research. My work there centered on advancing wake-word detection and audio event classification, with a focus on efficient architectural design, domain generalization, and few-shot open-set recognition. My current research interests lies at the general machine learning/ deep learning, with a particular emphasis on constructing large-scale models from a multi-modal perspective.

PUBLICATIONS & PATENTS

Please refer to Publications & Patents page.

Honors

Alexa Perceptual Technology Award Q3/Q4 2023, Recognized for contribution to Large Language Models that raise the functional bar. Awarded by Amazon in Feb 2024.

Byeonggeun Kim

Challenge Winner, IEEE AASP Challenge on Detection and Classification of Acoustic Scenes and Events (DCASE) Challenge, DCASE Challenge, 2021 [Results]

Byeonggeun Kim, Seunghan Yang, Jangho Kim, Simyung Chang

Academic Services

Session Chair: INTERSPEECH 2022

Reviewer: {INTERSPEECH, ICASSP, CVPR}

EDUCATION

M.S. Feb 2016 - Feb 2018

Korea Advanced Institute of Science and Technology (KAIST)

Electrical Engineering

Dissertion: Knowledge base completion with translation based embeddings and negative sampling method

Advisor: Soo-Young Lee

B.S. Feb 2009 - Feb 2016

Korea Advanced Institute of Science and Technology (KAIST)

Electrical Engineering

Double major in Business

Mandatory Military Service

Sergeant Jun 2010 - Mar 2012

Korea Augmentation To the Unitied States Army (KATUSA)

High School. Mar, 2006 - Feb, 2009

Korea Science Academy of KAIST (KSA)